Limits...
Mapping condition-dependent regulation of lipid metabolism in Saccharomyces cerevisiae.

Jewett MC, Workman CT, Nookaew I, Pizarro FA, Agosin E, Hellgren LI, Nielsen J - G3 (Bethesda) (2013)

Bottom Line: Here, we map condition-dependent regulation controlling lipid metabolism in Saccharomyces cerevisiae by measuring 5636 mRNAs, 50 metabolites, 97 lipids, and 57 (13)C-reaction fluxes in yeast using a three-factor full-factorial design.To query this network, we developed integrative methods for correlation of multi-omics datasets that elucidate global regulatory signatures.Our data highlight many characterized regulators of lipid metabolism and reveal that sterols are regulated more at the transcriptional level than are amino acids.

View Article: PubMed Central - PubMed

Affiliation: Center for Microbial Biotechnology, DTU Systems Biology, Technical University of Denmark, DK-2800 Kongens Lyngby, Denmark.

ABSTRACT
Lipids play a central role in cellular function as constituents of membranes, as signaling molecules, and as storage materials. Although much is known about the role of lipids in regulating specific steps of metabolism, comprehensive studies integrating genome-wide expression data, metabolite levels, and lipid levels are currently lacking. Here, we map condition-dependent regulation controlling lipid metabolism in Saccharomyces cerevisiae by measuring 5636 mRNAs, 50 metabolites, 97 lipids, and 57 (13)C-reaction fluxes in yeast using a three-factor full-factorial design. Correlation analysis across eight environmental conditions revealed 2279 gene expression level-metabolite/lipid relationships that characterize the extent of transcriptional regulation in lipid metabolism relative to major metabolic hubs within the cell. To query this network, we developed integrative methods for correlation of multi-omics datasets that elucidate global regulatory signatures. Our data highlight many characterized regulators of lipid metabolism and reveal that sterols are regulated more at the transcriptional level than are amino acids. Beyond providing insights into the systems-level organization of lipid metabolism, we anticipate that our dataset and approach can join an emerging number of studies to be widely used for interrogating cellular systems through the combination of mathematical modeling and experimental biology.

Show MeSH

Related in: MedlinePlus

Integrative method for correlation of multi-omics datasets reveals systems-level regulatory signatures. Correlation networks for ergosterol (A and B) and phosphatidylinositol (C and D) show first (green highlight) and second (blue highlight) significantly linked gene neighbors. (A and C) Genes in small white boxes were not identified as significantly correlated to ergosterol or phosphatidylinositol, but they are represented as “connector” nodes between metabolites. TFs implicated by the enrichment analysis are shown. Co-regulated gene neighborhood networks from (A and C) were expanded to include genes and metabolites necessary to perform the metabolic transformations indicated (B and D). This provides a more integrated perspective of cellular regulation. Measurement ratios for aerobic vs. anaerobic conditions were visualized with a log2 color bar, and the color of each node border represents the log10(P value). See node and edge color key. Gray indicates the lack of a measurement for that node.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3815060&req=5

fig7: Integrative method for correlation of multi-omics datasets reveals systems-level regulatory signatures. Correlation networks for ergosterol (A and B) and phosphatidylinositol (C and D) show first (green highlight) and second (blue highlight) significantly linked gene neighbors. (A and C) Genes in small white boxes were not identified as significantly correlated to ergosterol or phosphatidylinositol, but they are represented as “connector” nodes between metabolites. TFs implicated by the enrichment analysis are shown. Co-regulated gene neighborhood networks from (A and C) were expanded to include genes and metabolites necessary to perform the metabolic transformations indicated (B and D). This provides a more integrated perspective of cellular regulation. Measurement ratios for aerobic vs. anaerobic conditions were visualized with a log2 color bar, and the color of each node border represents the log10(P value). See node and edge color key. Gray indicates the lack of a measurement for that node.

Mentions: Finally, we built correlation networks for ergosterol and phosphatidylinositol based on the overlap between first and second gene neighbors from the co-regulated gene neighborhoods and the genome-scale metabolic map (Figure 7). The key idea was to identify areas of metabolism that are closely connected with significant and coordinated response to genetic or environmental perturbations. For ergosterol, we identified 76 significant gene expression level–ergosterol interactions in the correlation analysis (Figure S23) and mapped 102 genes linked to ergosterol in the genome scale metabolic model. By taking the intersection of these two groups, we identified nine metabolites that were both linked as a first and second gene neighbors to ergosterol in the metabolic map and were significantly correlated (e.g., ergosterol → ARE1 → acylCoA → FAA1). This topological map was then integrated with the TF-regulated modules implicated in the enrichment analysis to provide a global regulatory picture (Figure 7A). One TF, Upc2p, is known to activate the expression of sterol biosynthetic genes in sterol-depleted cells (Espenshade and Hughes 2007; Gaspar et al. 2007) and could be directly linked to Erg2p and Aus1p. The identification of Upc2p as significant when comparing aerobic vs. anaerobic conditions validates our undirected approach. Notably, our results suggest that Upc2p functions by controlling the enzyme levels at the transcriptional level and that this results in altered fluxes toward ergosterol. This points to Upc2p playing a similar function as the SREBP-1 transcription factor in mammals (Nielsen 2009; Nookaew et al. 2010). Our analysis also identified a significant link between ergosterol and 1-acyl-sn-glycerol-3-phosphate acyltransferase (SLC1), which is responsible for the synthesis of phosphatidic acid (PA). PA is the central precursor for glycerophospholipids, DAG, and TAG, and it is also a signaling lipid and key transcriptional regulator of lipid biosynthesis (Carman and Henry 2007). Our correlation network from the co-regulated gene neighborhoods was expanded to provide context within the scope of the genome-scale metabolic map (Figure 7B). Follow-up studies are expected to use the systems interactions here to bring new understanding to lipid metabolism.


Mapping condition-dependent regulation of lipid metabolism in Saccharomyces cerevisiae.

Jewett MC, Workman CT, Nookaew I, Pizarro FA, Agosin E, Hellgren LI, Nielsen J - G3 (Bethesda) (2013)

Integrative method for correlation of multi-omics datasets reveals systems-level regulatory signatures. Correlation networks for ergosterol (A and B) and phosphatidylinositol (C and D) show first (green highlight) and second (blue highlight) significantly linked gene neighbors. (A and C) Genes in small white boxes were not identified as significantly correlated to ergosterol or phosphatidylinositol, but they are represented as “connector” nodes between metabolites. TFs implicated by the enrichment analysis are shown. Co-regulated gene neighborhood networks from (A and C) were expanded to include genes and metabolites necessary to perform the metabolic transformations indicated (B and D). This provides a more integrated perspective of cellular regulation. Measurement ratios for aerobic vs. anaerobic conditions were visualized with a log2 color bar, and the color of each node border represents the log10(P value). See node and edge color key. Gray indicates the lack of a measurement for that node.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3815060&req=5

fig7: Integrative method for correlation of multi-omics datasets reveals systems-level regulatory signatures. Correlation networks for ergosterol (A and B) and phosphatidylinositol (C and D) show first (green highlight) and second (blue highlight) significantly linked gene neighbors. (A and C) Genes in small white boxes were not identified as significantly correlated to ergosterol or phosphatidylinositol, but they are represented as “connector” nodes between metabolites. TFs implicated by the enrichment analysis are shown. Co-regulated gene neighborhood networks from (A and C) were expanded to include genes and metabolites necessary to perform the metabolic transformations indicated (B and D). This provides a more integrated perspective of cellular regulation. Measurement ratios for aerobic vs. anaerobic conditions were visualized with a log2 color bar, and the color of each node border represents the log10(P value). See node and edge color key. Gray indicates the lack of a measurement for that node.
Mentions: Finally, we built correlation networks for ergosterol and phosphatidylinositol based on the overlap between first and second gene neighbors from the co-regulated gene neighborhoods and the genome-scale metabolic map (Figure 7). The key idea was to identify areas of metabolism that are closely connected with significant and coordinated response to genetic or environmental perturbations. For ergosterol, we identified 76 significant gene expression level–ergosterol interactions in the correlation analysis (Figure S23) and mapped 102 genes linked to ergosterol in the genome scale metabolic model. By taking the intersection of these two groups, we identified nine metabolites that were both linked as a first and second gene neighbors to ergosterol in the metabolic map and were significantly correlated (e.g., ergosterol → ARE1 → acylCoA → FAA1). This topological map was then integrated with the TF-regulated modules implicated in the enrichment analysis to provide a global regulatory picture (Figure 7A). One TF, Upc2p, is known to activate the expression of sterol biosynthetic genes in sterol-depleted cells (Espenshade and Hughes 2007; Gaspar et al. 2007) and could be directly linked to Erg2p and Aus1p. The identification of Upc2p as significant when comparing aerobic vs. anaerobic conditions validates our undirected approach. Notably, our results suggest that Upc2p functions by controlling the enzyme levels at the transcriptional level and that this results in altered fluxes toward ergosterol. This points to Upc2p playing a similar function as the SREBP-1 transcription factor in mammals (Nielsen 2009; Nookaew et al. 2010). Our analysis also identified a significant link between ergosterol and 1-acyl-sn-glycerol-3-phosphate acyltransferase (SLC1), which is responsible for the synthesis of phosphatidic acid (PA). PA is the central precursor for glycerophospholipids, DAG, and TAG, and it is also a signaling lipid and key transcriptional regulator of lipid biosynthesis (Carman and Henry 2007). Our correlation network from the co-regulated gene neighborhoods was expanded to provide context within the scope of the genome-scale metabolic map (Figure 7B). Follow-up studies are expected to use the systems interactions here to bring new understanding to lipid metabolism.

Bottom Line: Here, we map condition-dependent regulation controlling lipid metabolism in Saccharomyces cerevisiae by measuring 5636 mRNAs, 50 metabolites, 97 lipids, and 57 (13)C-reaction fluxes in yeast using a three-factor full-factorial design.To query this network, we developed integrative methods for correlation of multi-omics datasets that elucidate global regulatory signatures.Our data highlight many characterized regulators of lipid metabolism and reveal that sterols are regulated more at the transcriptional level than are amino acids.

View Article: PubMed Central - PubMed

Affiliation: Center for Microbial Biotechnology, DTU Systems Biology, Technical University of Denmark, DK-2800 Kongens Lyngby, Denmark.

ABSTRACT
Lipids play a central role in cellular function as constituents of membranes, as signaling molecules, and as storage materials. Although much is known about the role of lipids in regulating specific steps of metabolism, comprehensive studies integrating genome-wide expression data, metabolite levels, and lipid levels are currently lacking. Here, we map condition-dependent regulation controlling lipid metabolism in Saccharomyces cerevisiae by measuring 5636 mRNAs, 50 metabolites, 97 lipids, and 57 (13)C-reaction fluxes in yeast using a three-factor full-factorial design. Correlation analysis across eight environmental conditions revealed 2279 gene expression level-metabolite/lipid relationships that characterize the extent of transcriptional regulation in lipid metabolism relative to major metabolic hubs within the cell. To query this network, we developed integrative methods for correlation of multi-omics datasets that elucidate global regulatory signatures. Our data highlight many characterized regulators of lipid metabolism and reveal that sterols are regulated more at the transcriptional level than are amino acids. Beyond providing insights into the systems-level organization of lipid metabolism, we anticipate that our dataset and approach can join an emerging number of studies to be widely used for interrogating cellular systems through the combination of mathematical modeling and experimental biology.

Show MeSH
Related in: MedlinePlus